JPH02188079A - Motion detection system - Google Patents

Abstract

PURPOSE:To perform motion detection speedily with simple circuit constitution by comparing block characteristic information with past block characteristic information which is at spatially the same position and detecting whether or not a corresponding block is in motion. CONSTITUTION:The in-block mean value (m) of a data separating circuit 14 is applied to a block mean memory 18 and a motion decision circuit 20. The motion decision circuit 20 compares the in-block mean value which is stored in the block main memory 18 and at the same block position in a last frame with the current in-block mean value from the data separating circuit 14 and sets a proper threshold value as, for example, the difference to decide whether or not the block is in motion. The decision result of the motion decision circuit 20 is represented as the specific number of bits, which is written in a block motion memory 22. Thus, decoded data is written in a frame memory 28 and motion information is written in the block motion memory 22.

Description

[Detailed description of the invention] [Industrial application fields] The present invention relates to a motion detection method for image data.

[Prior Art] A DC separation vector quantization method is known as an encoding method for efficiently encoding image data in units of blocks. This takes advantage of the fact that the intra-block amplitude distribution of the image signal after subtracting the intra-block average value is concentrated around O, vector quantizes the image signal after subtracting the average value, and converts its index data and average value data. It is something that is transmitted.

In such a block encoding method, in the transmission path,
If data loss occurs in information representing the properties of a block, such as the intra-block average value, it becomes impossible to decode each block, resulting in a large visual disturbance. Therefore, there is a need to reduce this visual disturbance by block-by-block interpolation.

[Problem to be solved by the invention] When performing block-by-block interpolation, in a static area of an image,
Interframe interpolation using pixel values of past frames is effective and can provide extremely high visual improvement effects. Therefore, the amount of motion of nearby decodable blocks is detected, and the amount of motion of undecodable blocks is estimated based on the detection result. If the block is determined to be a stationary block, the undecodable block is interpolated using interframe interpolation. . When decoded pixel values are used to detect motion of a block, motion detection cannot be performed until decoding of neighboring plots is completed, which is disadvantageous when decoding by vibrating processing is performed. Furthermore, comparing all pixels for the purpose of interpolation when data is missing is wasteful in terms of calculation time and calculation circuitry.

Therefore, it is an object of the present invention to present a motion detection method that solves these problems.

[Means for Solving the Problems] The motion detection method according to the present invention divides image data into blocks each consisting of a plurality of pixels, and calculates the pixel values in each block by
In a system that encodes image data based on block characteristic information indicating the characteristics of the block and transmits image data using the block characteristic information and the encoded information of each pixel value, the block characteristic information is encoded based on the block characteristic information of a past block at the same spatial location. The feature is that the presence or absence of movement of the block is detected by comparing it with characteristic information.

[Operation] With the above means, the presence or absence of movement is determined on a block-by-block basis, so it can be performed quickly with a simple circuit configuration, and is suitable for vibration line processing.

[Example] Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 shows a configuration block diagram of an embodiment in which the present invention is applied to a decoding device in an image data transmission system using a DC separation vector quantization method. Reference numeral 10 denotes an input terminal for transmission data from a transmission path, into which encoded data compressed by DC separation vector quantization is input. 12 is a block missing determination circuit that determines whether or not data loss has occurred in each block; 14 is a circuit that subtracts received data from the intra-block average value m (block characteristic information) and subtracts the intra-block average value m from each pixel value within the block; a data separation circuit that separates the obtained value into vector quantized data, that is, vector quantization index i (intra-block pixel information);
17 is an inverse vector quantization circuit that inverse vector quantizes the vector quantization index i, and 17 is an addition circuit that adds the intra-block average value m from the data separation circuit 14 to the output of the inverse vector quantization circuit 16.

18 is a block average memory that stores the intra-block average value for one screen; 20 is a motion determination circuit that determines the presence or absence of movement in each block by comparing the current and past intra-block average values m; 22 is a motion determination circuit; A block motion memory 24 that stores the determination result of the determination circuit 22 determines an interpolation method by estimating the presence or absence of movement of the missing block from the motion information of neighboring blocks when a block is missing in units of blocks. An interpolation method determination circuit 26 is an interpolation method determination circuit 24
The adaptive interpolation circuit 27 performs inter-frame interpolation for stationary blocks and intra-frame interpolation for moving blocks according to the determination of . An adaptive intra-block average interpolation circuit performs interpolation using a similar interpolation method; 28 is a frame memory that stores decoded data from the adder circuit 17 and interpolated data from the adaptive interpolation circuit 26; and 30 is an output terminal.

The operation shown in FIG. 1 will be explained. Transmission data input from the transmission path to the input terminal 10 is input to the block missing determination circuit 12, and the block missing determining circuit 12 determines whether or not data is missing in units of blocks, and if data is missing, it outputs a block missing signal. e is applied to the interpolation method determining circuit 24. The data of the block without missing data is applied to the data separation circuit 14 and separated into the intra-block average value m and the vector quantization index i. The inverse vector quantization circuit 16 inverse vector quantizes the vector quantization index i, and the addition circuit 17 adds the intra-block average value m from the data separation circuit 14 to the inverse vector quantized data. As a result, decoded pixel data is obtained.

The intra-block average value m of the data separation circuit 14 is also applied to the block average memory 18 and the motion determination circuit 20,
The motion determination circuit 20 compares the intra-block average value at the same block position one frame before, which is stored in the block average memory 18, with the current intra-block average value from the data separation circuit 14, and calculates the difference, for example. The presence or absence of block movement is determined by setting an appropriate threshold. The determination result of the motion determination circuit 20 is expressed by a predetermined number of bits, and is written into the block motion memory 22.

The data stored in the block average memory 18 is rewritten with the intra-block average value of the current frame.

Through the above processing, decoded data is written to the frame memory 28 for blocks without data loss,
Motion information is written to block motion memory 22. Regarding the missing block, the interpolation method determining circuit 24 refers to the data stored in the block motion memory 22, estimates the motion information of the missing block from the motion information of neighboring blocks among the decoded blocks, and uses the estimation result as the missing block. While writing the block motion information to the block motion memory 22,
The signal is applied to the adaptive interpolator 26 and the adaptive block average interpolator 27 .

Adaptive interpolation circuit 26 selects interframe interpolation for static blocks and intraframe interpolation for moving blocks, forms interpolated data using data stored in frame memory 28, and writes it to frame memory 28. In addition,
For interframe interpolation, the frame memory 28
Since the image data of the previous frame remains in the memory location corresponding to the missing block, inter-frame interpolation can be realized by prohibiting data rewriting. Regarding the intra-frame sleeve, data of all pixels of the missing block is interpolated using data of neighboring decoded blocks.

On the other hand, the adaptive intra-block average interpolation circuit 27 interpolates the intra-block average value of the missing block using the same interpolation method as the adaptive interpolation circuit 26, and stores the interpolated data in the block average memory 1.
Write in 8. Thereby, even if two or more frames of blocks at the same position are missing consecutively, motion determination can be performed using interpolated data.

In this embodiment, a decoding device for encoded data using DC separation vector quantization is taken as an example, but the present invention also provides a block representing the properties of the block (block characteristic information) and information of each pixel in the block (block It can be widely applied to image transmission systems that express image information using internal pixel information). As block characteristic information, in addition to the average value,
The minimum value, maximum value, dynamic range, variance, median value, etc., or a combination of these may be used. In addition to the vector quantization index after DC separation, the intra-block pixel information includes an index obtained by normalizing the intra-block pixel value after minimum value removal by the intra-block dynamic range and quantizing it in -dimensional or multidimensional, or orthogonal It may also be a conversion coefficient or the like. In this case, the block characteristic information may include the minimum value and maximum value, the minimum value and dynamic range, etc. As an example of motion determination, the minimum value and maximum value (dynamic range) can be calculated based on the values in the previous frame. compared to
A configuration may be considered in which the presence or absence of movement is determined by a linear combination of the results. With this configuration, the number of parameters for determination increases, so more accurate motion determination can be performed.

[Effects of the Invention] As can be easily understood from the above description, according to the present invention, motion detection can be performed simply and at high speed using a method suitable for pipeline processing.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the present invention. 10: Input terminal 12 Niblock missing determination circuit 14: Data separation circuit 16: Inverse vector quantization circuit 18 Niblock average memory 20: Motion determination circuit 22 Niblock motion memory 24: Interpolation method determination circuit 26: Adaptive interpolation circuit 27: Adaptation Intra-block average interpolation circuit 28: Frame memory 30: Output terminal

Claims (1)

[Claims] The image data is divided into blocks consisting of multiple pixels, the pixel values in each block are encoded based on block characteristic information indicating the characteristics of the block, and the image data is created using the block characteristic information and the encoding information of each pixel value. In a data transmission system, a motion detection method is characterized in that the characteristic information of the block is compared with the characteristic information of past blocks at the same spatial location to detect the presence or absence of movement of the block.